Upload config.json
Browse files- config.json +43 -0
config.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/hf-trainer/longt5-large-scitldr/checkpoint-24",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LongT5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"d_ff": 2816,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 1024,
|
| 9 |
+
"decoder_start_token_id": 0,
|
| 10 |
+
"dense_act_fn": "gelu_new",
|
| 11 |
+
"dropout_rate": 0.1,
|
| 12 |
+
"early_stopping": true,
|
| 13 |
+
"encoder_attention_type": "transient-global",
|
| 14 |
+
"encoder_no_repeat_ngram_size": 4,
|
| 15 |
+
"eos_token_id": 1,
|
| 16 |
+
"feed_forward_proj": "gated-gelu",
|
| 17 |
+
"global_block_size": 16,
|
| 18 |
+
"initializer_factor": 1.0,
|
| 19 |
+
"is_encoder_decoder": true,
|
| 20 |
+
"is_gated_act": true,
|
| 21 |
+
"layer_norm_epsilon": 1e-06,
|
| 22 |
+
"length_penalty": 0.6,
|
| 23 |
+
"local_radius": 127,
|
| 24 |
+
"max_length": 62,
|
| 25 |
+
"min_length": 11,
|
| 26 |
+
"model_type": "longt5",
|
| 27 |
+
"n_positions": 4096,
|
| 28 |
+
"no_repeat_ngram_size": 3,
|
| 29 |
+
"num_beams": 4,
|
| 30 |
+
"num_decoder_layers": 24,
|
| 31 |
+
"num_heads": 16,
|
| 32 |
+
"num_layers": 24,
|
| 33 |
+
"output_past": true,
|
| 34 |
+
"pad_token_id": 0,
|
| 35 |
+
"relative_attention_max_distance": 128,
|
| 36 |
+
"relative_attention_num_buckets": 32,
|
| 37 |
+
"repetition_penalty": 3.5,
|
| 38 |
+
"tie_word_embeddings": false,
|
| 39 |
+
"torch_dtype": "float32",
|
| 40 |
+
"transformers_version": "4.21.1",
|
| 41 |
+
"use_cache": false,
|
| 42 |
+
"vocab_size": 32100
|
| 43 |
+
}
|